BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 1079-1085
Vo1.173, No. 3,1990 December 31, 1990
ISOLATION AND CHARACTERIZATION OF A GALACTOSE-SPECIFIC CARBOHYDRATE BINDING PROTEIN FROM HUMAN LYMPHOBLASTIC CELLS Linda G. Baum* ,1 and Peter Cresswell 2 1Department of Pathology, UCLA School of Medicine, Los Angeles, CA 90024 2Department of Microbiology and Immunology, Duke University School of Medicine, Durham, NC 27710 Received November 5, 1990
Summary A carbohydrate binding protein of Mr = 32,000 (CBP 32) has been isolated from detergent extracts of human B and T lymphoblastoid cells. CBP 32 binds specifically to glycoproteins containing asparagine-linked complex oligosaccharides, and can be eluted from a fetuin affinity matrix by #-lactose. Binding is not thiol dependent, nor are divalent cations necessary for binding. Native CBP 32 appears to exist as a monomer, with a pl of 8.2. Purified CBP 32 can bind detergent, as shown by chargeshift electrophoresis, and thus appears to be an integral membrane protein. ~1990Aoademio Pressl
Inc.
The role of carbohydrate binding proteins in lymphocyte function has recently been the subject of intense investigation (1-4). Structural similarities can be found among these lectins (3,5,6), some of which have similar function, such as control of cellular adhesion and migration. The best characterized lymphocyte lectin is MEL-14, a murine lymphocyte homing receptor (3,7,8). Soluble galactose-binding lectins (galaptins), postulated to mediate cell adhesion, have been isolated from murine thymus (4,9) and human spleen (10). In addition, membrane-bound lectins with specificity for galactose, mannose and N-acetyl-galactosamine have been described in murine lymphoid cells (11,12), although their function is unknown.
Our laboratory has described a cell-surface carbohydrate binding activity in human B and T lymphoblastoid cell lines and in activated, but not resting, peripheral blood lymphocytes (13). This lectin was identified by binding water-soluble glycoprotein micelles to the lymphoblastoid cells; binding of the micelles was inhibited by asparagine-
*To whom correspondence should be addressed.
1079
0006-291X/90 $1.50 Copyright © 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
Vol. 173, No. 3, 1990
. BIOCHEMICALAND BIOPHYSICALRESEARCHCOMMUNICATIONS
linked glycopeptides derived from fetuin. Additional studies revealed this binding was also inhibitable by/~-galactosides, such as ~-Iactose and thiodigalactoside. 1
We report the isolation of a/3-galactoside specific lectin from detergent extracts of human B and T lymphoblastoid cells. This protein appears to be an integral membrane protein, and thus may be responsible for the cell surface carbohydrate binding activity previously described.
Materials and Methods Cells: The B and T lymphoblastoid cell lines TRal (B), Swei (B) and CEM(T) were cultured in Iscove's medium with 3% fetal calf serum (Gibco), with gentamycin, 10/Jg/ml (Schering). Biosynthetic labeling: Prior to labeling, 1-2 x 107 cells were pre-incubated for one hour at 37 ° in methionine-free RPMI 1640 medium with 3% dialyzed fetal calf serum. The cells were labeled with 1 mCi [35S]-methionine (NEN) in 2 ml fresh methionine-free medium for four hours at 37°. After labeling, the cells were washed three times in 10 mM Tris, 150 mM NaCI, 3.0 mM sodium azide, pH 7.4 (buffer A) and solubilized in 1 ml of 4% (w/v) polyoxyethylene 9 lauryl ether (£;12E9, Sigma) in buffer A with 0.5 mM tosyllysine chloromethyl ketone (TLCK, Sigma) and 1 mM phenylmethylsulfonyl fluoride (PMSF, Sigma) for 30 minutes on ice. Extracts were centrifuged in a Beckman Microfuge at 500 g for 30 seconds to pellet nuclei, then at 12,000 g for 5 minutes to pellet residual detergent insoluble material. Affinity purification: Fetuin (Sigma) was coupled to Biogel A15M and deoxyribonuclease (DNase I from bovine pancreas, Sigma) was coupled to Sephacryl $200 as described (14) at approximately 10 mg/ml of resin. For purification from [35S]labeled cells, extracts (1 ml) were first rotated with 300 pl of DNase-S200 for 2 hours at 4° to reduce non-specific binding. The supernatants were then rotated with 300/A fetuin-A15M, for eight hours at 4 °. After extensive washing with 0.5% (w/v) sodium deoxycholate in 10 mM Tris, 150 mM NaCI, 3 mM sodium azide, pH 8.0 (buffer B), followed by 0.1% (w/v) C12E 9 in buffer A, bound material was eluted with 150 mM/3lactose (Sigma), 0.1% C12E 9 in 1 ml buffer A for eight hours at 4°. The eluate was either used directly in subsequent procedures, or dialyzed versus buffer A at 4° and precipitated at -20° with 5 volumes of acetone containing 0.1% (X/v) HCI. For large scale purification of unlabeled material, 2-6 x 10= cells were extracted with 4% (w/v) C12E9 in buffer A (25 ml/109 cells) with PMSF and TLCK as above. Extracts were centrifuged at 1200 x g for 15 minutes at 4° in a Sorvall RC-2 centrifuge to pellet nuclei, and the supernatants were centrifuged at 100,000 x g for 90 minutes in a Beckman model L5-65 ultracentrifuge. The supernatants were applied to a 8 ml column of DNase-S200 and a 20 ml column of fetuin-A15M in tandem at 4°. After washing the fetuin-A15M column extensively with buffer B, bound material was eluted with 150 mM /3-lactose in buffer B. Fractions were pooled, concentrated by ultrafiltration using a PM 10 membrane (Amicon) and frozen at -70°. Approximate yield was 50-100/~g per 109 cells. For iodination, 10/Jg of isolated protein was labeled with 1 mCi Bolton-Hunter reagent (NEN). 1Baum, LG and P Cresswell, unpublished results. 1080
Vol. 173, No. 3, 1 9 9 0
BIOCHEMICAL AND BIOPHYSICALRESEARCHCOMMUNICATIONS
Gel electrophoresis: One dimension SDS-PAGE was performed under reducing conditions (0.1% (v/v) 2-mercaptoethanol) as described by Maizel (15). Two dimension SDS-PAGE, with the addition of St. aureus V-8 protease (20/Jg/ml, Sigma) in the second dimension, was performed essentially as described by Cleveland (16). All SDS-PAGE gels were 10.5% (w/v) acrylamide. Isoelectric focussing was performed over a gradient of 3-10 according to the method of O'Farrell (17) modified so that focusing was performed on a 10 cm slab gel using a 2117 Multiphor slab gel apparatus (LKB). Standard proteins (Pharmacia) were run parallel to the sample. For charge-shift electrophoresis (18), 10/JI samples were applied to wells in the middle of 7.5 cm 1% agarose slab gels, in the appropriate detergent-containing buffers. Samples were subjected to electrophoresis for 80 minutes at 30 volts. Results Detergent extracts of one T and two B lymphoblastoid cells were applied to an affinity matrix of fetuin-A15M. Elution with 150 mM/3-lactose yielded the material shown in Figure 1, while incubation with the following saccharides were ineffective in eluting any material bound to the fetuin-A15M: galactose, N-acetyl galactosamine, mannose, glucose, N-acetylglucosamine, and xylose. In all three cell lines, the major carbohydrate binding protein eluted by /3-1actose has a molecular weight of 32,000 (CBP 32) by SDS-PAGE analysis. In addition, proteins of Mr = 29,000 (Swei, CEM) and Mr = 25,000 (TRal) are variably seen; these are presumed to be proteolytic products of CBP 32 (see below). [35S]-Iabelled preparations often contain material of Mr = 45,000 (e.g., CEM), which appears to be actin, by pl and differential mobility under reducing versus non-reducing conditions, as compared with purified actin (data not shown). To determine whether 29 kD material seen in Figure 1 is a proteolytic product of CBP 32, affinity purified unlabelled material containing both species was first subjected to
44> 40>
14~
CEM
Swei
TRal
Figure 1. Affinity purified CBP 32 from CEM(T), Swei(B) and TRaI(B) lymphoblastoid cells. Detergent extracts of [35S] labelled (CEM) or unlabeled (Swei,TRal) lymphoblastoid cells were applied to a fetuin-A15M column and eluted with 150 mM /3-lactose. Swei and TRal lanes were silver stained. 1081
Vol. 173, No. 3, 1990
BIOCHEMICAL AND BIOPHYSICALRESEARCH COMMUNICATIONS
67~,
B
44~. 40~..
14~
®
®
,23
Figure 2. The 32 kD and 29 kD species isolated by affinity chromatography appear to be related. Unlabelled material isolated from Swei cells was applied to 10 cm tube gels and subjected to SDS-PAGE. The tube gels were applied to slab gels after equilibration in sample buffer with or without the addition of 2/~g/ml S. aureus V-8 protease. Slab gels were silver stained. The appearance of only one major spot after digestion suggests that the protein has a protease-resistant core. Panel A, no protease added. Panel B, 2 pg/ml protease added to the sample buffer. Figure ~1"251]labeled CBP 32 isolated from Swei cells was reapplied to fetuin A15M. Both CBP 32 and the 29 kD proteolytic product were re-eluted with 150 mM /~-Iactose (lane 1) but neither 2 mM dithiothreitol in buffer B (lane 2) or 5 mM EDTA in buffer B (lane 3) eluted bound material from the affinity matrix.
electrophoresis on SDS-PAGE tube gels, and then applied to SDS-PAGE slab gels in the presence of St. aureus V-8 protease (20 pg/ml) (Figure 2). The peptide fragments from the 32 kD(a) and 29 kD(b) bands resolved on silver staining appear to be related. Both polypeptides were digested to one major spot (a' derived from a), approximately 8 kD less than the undigested precursors, and one minor spot (b, derived from b). The fragments derived from the 32 kD and 29 kD bands have the same spatial relationship on the gel as the intact molecules. This implies that both molecules lose the same peptide upon proteolysis. The 25 kD protein seen in Figure 1 appears primarily after 1082
Vol. 173, No. 3, 1990
BIOCHEMICALAND BIOPHYSICAL RESEARCH COMMUNICATIONS
i~i!i!i!i!~ li !,!i ii!i~~!
81x5_ 84,5 8.15 -
~
+
735 6.85
635-
Q
stds
CBP 32
®+
~
2
3
4
Figure 4. Isoelectric focussing of CBP 32. The sample was applied to the middle of a 3:10 gradient slab gel. Proteins were detected by Coomassie Blue staining. Isoelectric point standards: lentil lectin, basic band, 8.65; lentil lectin, middle band, 8.45; lentil lectin, acidic band, 8.15; horse myoglobin, basic band, 7.35; horse myoglobin, acidic band, 6.85; human carbonic anhydrase, 6.55. Figure ['1"251]labeled CBP 32 isolated from Swei cells was subjected to charge-shift electrophoresis. In sodium deoxycholate, CBP 32 migrated towards the anode (+); in cetyltrimethyl ammonium bromide, CBP 32 migrated towards the cathode (-); in Triton X100, CBP 32 did not migrate from the origin. Lane 1, CBP 32; lane 2, soluble F(ab, )2 fragments of rabbit immunoglobulin; lane 3, glycoproteins from TRal (mostly integral membrane proteins); lane 4, purified HLA class II molecules (a heterodimer of integral membrane glycoproteins). repeated freeze-thaw cycles; the irregular appearance of this product has prevented similar analysis. Unlike the soluble (S-type) lectins (5,19), the binding activity of CBP 32 did not appear to be thiol dependent, since the inclusion of reducing agents in extraction and elution buffers did not affect yields (data not shown). Figure 3 shows that the addition of dithiothreitol does not dissociate bound CBP 32, indicating that maintenance of an active binding site does not depend on intramolecular disulphide bonds. In addition, divalent cations do not appear to be essential for binding, as the presence of EDTA does not dissociate bound lectin (Figure 3). Native CBP 32 appears to exist as a monomer, since the protein could not be cross-linked by the bifunctional cross-linking reagent dithiobis (propionic acid N-hydroxysuccinimide ester) (data not shown). The isoelectric point of affinity-purified material was determined over a pH range of 3-10 (Figure 4). The proteins resolved into 3 distinct bands. This material was the 1083
Vol. 173, No. 3, 1 9 9 0
BIOCHEMICALAND BIOPHYSICALRESEARCHCOMMUNICATIONS
identical preparation used in Figure la (Swei); on the basis of staining intensity, the major band at 8.2 corresponds to the 32 kD protein. The two minor bands, at 8.5 and 8.7, thus represent the 29 kD and 25 kD material seen in Figure 1. This relationship suggests that the 25 kD protein is, like the 29 kD protein, a proteolytic product derived from CBP 32. While some soluble proteins may bind detergent, the majority of proteins capable of interacting with detergents are integral membrane proteins, which associate with detergent molecules via their hydrophobic transmembrane regions. To investigate whether CBP 32 is an integral membrane protein, the ability of the protein to bind detergent was examined by charge-shift electrophoresis. As shown in Figure 5, the direction of CBP 32 migration in 1% agarose gels changes according to the charge of the detergent used in the buffer solutions. In the top panel, CBP 32, equilibrated in the anionic detergent sodium deoxycholate, migrates towards the anode, while, in the middle panel, CBP 32 migrates towards the cathode in the presence of cetyltrimethylammonium bromide, a cationic detergent. CBP 32 does not migrate from the loading point in the neutral detergent Triton X-100 (bottom panel). These results indicate that CBP 32 has the ability to bind detergent, and thus is likely an integral membrane protein. Discussion Based on the previous observation that binding of glycoprotein micelles to activated lymphocytes and to B and T lymphoblastoid cells was inhibitable by -galactosides, we examined detergent extracts of these cells for carbohydrate binding proteins. Affinity chromatography of cell extracts on fetuin-A15M yielded one major carbohydrate binding (CBP 32), as well as two smaller proteins which appear to be proteolytic products of CBP 32. While a number of galactose specific lectins have been described in a variety of mammalian tissues, many of these are S-type lectins (5,19,20); CBP 32 differs from this class of lectins in that binding is not thiol dependent. In addition, CBP 32 appears to be an integral membrane protein, based on charge-shift electrophoresis studies. Indeed, attempts to extract CBP 32 with lactose in the absence of detergent were unsuccessful. CBP 32 may have similarities to the other major class of carbohydrate binding proteins, the C-type lectins, which are a heterogeneous group of soluble and membrane bound proteins with a wide range of saccharide specificities, yet which share a common carbohydrate recognition domain (5,7). However, unlike the C-type lectins, CBP 32 activity does not appear to be calcium dependent, as the presence of calcium was not necessary for purification, nor did EDTA elute bound lectin from fetuin-A15M. Further characterization of CBP 32 will allow more extensive comparison with other C-type lectins. 1084
Vol. 173, No. 3, 1 9 9 0
BIOCHEMICAL AND BIOPHYSICALRESEARCHCOMMUNICATIONS
The role of CBP 32 in human lymphoblastoid cells is not known. Lectin-sugar interactions appear to be important for a variety of human and murine lymphocyte functions, such as homing to lymphoid tissue (7,8), localization and maturation of thymocytes (4,6,9) and T cell activation (21,22). The initial finding that glycoprotein micelles bind only to activated and transformed lymphocytes, but not to resting cells, suggested that a galactose specific lectin may be involved in immune responses (13). Additional studies will reveal whether the novel lectin described here, CBP 32, is involved in interactions among activated and transformed cells. Acknowledgments The authors thank Alan Payne for photography, and Janet Railsback for manuscript preparation. Re, fences 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
Hooghe, R.J., and Pink, J.R.L. (1985) Immunology Today 6, 180-181. Coombe, D.R., and Rider, C.C. (1989) Immunology Today 10, 289-291. Stoolman, L.M. (1989) Cell 56, 907-910. Levi, G., and Teichberg, V.I. (1985) Biochem J. 226, 379-384. Drickamer, K. (1988) J. Biol. Chem. 263, 9957-9560. Sharon, N., and Lis, H. (1989) Science 246, 227-234. Laskey, L.A., Singer, M.S., Yednock, T.A., Dowbenko, D., Fennie, C., Rodriguez, H., Nguyen, T., Stachel, S., and Rosen, S.D. (1989) Cell 56, 1045-1055. Geoffroy, J.S., and Rosen, S.D. (1989) J. Cell Biol. 109, 2463-2469. Levi, G., and Teichberg, V.I. (1983) Immunol. Lett. 7, 35-39. Lee, R.T., Ichikawa, Y., Allen, H.J., and Lee, Y.C. (1990) J. Biol. Chem. 265, 7864-7871. Keida, C.M.T., Bowles, D.J., Rand, A., and Sharon, N. (1978) FEBS Lett. 94, 391-396. Decker, J.M. (1980) Molecular Immunology 17, 803-808. Apgar, J.R., and Cresswell, P. (1982) Eur. J. Immunol. 12, 570-576. Axen, R., Porath, J., and Ernback, S. (1967) Nature 214, 1302-1304. Maizel, J.V. (1971) Methods in Virol. 5, 179-245. Cleveland, D.W., Fischer, S.G., Kirschner, M.W., and Laemmli, U.K. (1976) J. Biol. Chem. 252, 1102-1106. O'Farrell, P.H. (1975) J. Biol. Chem. 250, 4007-4021. Helenius, A., and Simons, K. (1977) Proc. Natl. Acad. Sci. 74, 529-532. Barondes, S. (1984) Science 223, 1259-1264. Sparrow, C.P., Leffler, H., and Barondes, S. (1987) J. Biol. Chem. 262, 73837390. Hart, G. (1982) J. Biol. Chem. 257, 151-158. Piller,F., Piller, V., Fox, R.I., and Fukuda, M. (1988) J. Biol. Chem. 263, 1514615150.
1085
Vo1.173, No. 3,1990 December 31, 1990
ISOLATION AND CHARACTERIZATION OF A GALACTOSE-SPECIFIC CARBOHYDRATE BINDING PROTEIN FROM HUMAN LYMPHOBLASTIC CELLS Linda G. Baum* ,1 and Peter Cresswell 2 1Department of Pathology, UCLA School of Medicine, Los Angeles, CA 90024 2Department of Microbiology and Immunology, Duke University School of Medicine, Durham, NC 27710 Received November 5, 1990
Summary A carbohydrate binding protein of Mr = 32,000 (CBP 32) has been isolated from detergent extracts of human B and T lymphoblastoid cells. CBP 32 binds specifically to glycoproteins containing asparagine-linked complex oligosaccharides, and can be eluted from a fetuin affinity matrix by #-lactose. Binding is not thiol dependent, nor are divalent cations necessary for binding. Native CBP 32 appears to exist as a monomer, with a pl of 8.2. Purified CBP 32 can bind detergent, as shown by chargeshift electrophoresis, and thus appears to be an integral membrane protein. ~1990Aoademio Pressl
Inc.
The role of carbohydrate binding proteins in lymphocyte function has recently been the subject of intense investigation (1-4). Structural similarities can be found among these lectins (3,5,6), some of which have similar function, such as control of cellular adhesion and migration. The best characterized lymphocyte lectin is MEL-14, a murine lymphocyte homing receptor (3,7,8). Soluble galactose-binding lectins (galaptins), postulated to mediate cell adhesion, have been isolated from murine thymus (4,9) and human spleen (10). In addition, membrane-bound lectins with specificity for galactose, mannose and N-acetyl-galactosamine have been described in murine lymphoid cells (11,12), although their function is unknown.
Our laboratory has described a cell-surface carbohydrate binding activity in human B and T lymphoblastoid cell lines and in activated, but not resting, peripheral blood lymphocytes (13). This lectin was identified by binding water-soluble glycoprotein micelles to the lymphoblastoid cells; binding of the micelles was inhibited by asparagine-
*To whom correspondence should be addressed.
1079
0006-291X/90 $1.50 Copyright © 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
Vol. 173, No. 3, 1990
. BIOCHEMICALAND BIOPHYSICALRESEARCHCOMMUNICATIONS
linked glycopeptides derived from fetuin. Additional studies revealed this binding was also inhibitable by/~-galactosides, such as ~-Iactose and thiodigalactoside. 1
We report the isolation of a/3-galactoside specific lectin from detergent extracts of human B and T lymphoblastoid cells. This protein appears to be an integral membrane protein, and thus may be responsible for the cell surface carbohydrate binding activity previously described.
Materials and Methods Cells: The B and T lymphoblastoid cell lines TRal (B), Swei (B) and CEM(T) were cultured in Iscove's medium with 3% fetal calf serum (Gibco), with gentamycin, 10/Jg/ml (Schering). Biosynthetic labeling: Prior to labeling, 1-2 x 107 cells were pre-incubated for one hour at 37 ° in methionine-free RPMI 1640 medium with 3% dialyzed fetal calf serum. The cells were labeled with 1 mCi [35S]-methionine (NEN) in 2 ml fresh methionine-free medium for four hours at 37°. After labeling, the cells were washed three times in 10 mM Tris, 150 mM NaCI, 3.0 mM sodium azide, pH 7.4 (buffer A) and solubilized in 1 ml of 4% (w/v) polyoxyethylene 9 lauryl ether (£;12E9, Sigma) in buffer A with 0.5 mM tosyllysine chloromethyl ketone (TLCK, Sigma) and 1 mM phenylmethylsulfonyl fluoride (PMSF, Sigma) for 30 minutes on ice. Extracts were centrifuged in a Beckman Microfuge at 500 g for 30 seconds to pellet nuclei, then at 12,000 g for 5 minutes to pellet residual detergent insoluble material. Affinity purification: Fetuin (Sigma) was coupled to Biogel A15M and deoxyribonuclease (DNase I from bovine pancreas, Sigma) was coupled to Sephacryl $200 as described (14) at approximately 10 mg/ml of resin. For purification from [35S]labeled cells, extracts (1 ml) were first rotated with 300 pl of DNase-S200 for 2 hours at 4° to reduce non-specific binding. The supernatants were then rotated with 300/A fetuin-A15M, for eight hours at 4 °. After extensive washing with 0.5% (w/v) sodium deoxycholate in 10 mM Tris, 150 mM NaCI, 3 mM sodium azide, pH 8.0 (buffer B), followed by 0.1% (w/v) C12E 9 in buffer A, bound material was eluted with 150 mM/3lactose (Sigma), 0.1% C12E 9 in 1 ml buffer A for eight hours at 4°. The eluate was either used directly in subsequent procedures, or dialyzed versus buffer A at 4° and precipitated at -20° with 5 volumes of acetone containing 0.1% (X/v) HCI. For large scale purification of unlabeled material, 2-6 x 10= cells were extracted with 4% (w/v) C12E9 in buffer A (25 ml/109 cells) with PMSF and TLCK as above. Extracts were centrifuged at 1200 x g for 15 minutes at 4° in a Sorvall RC-2 centrifuge to pellet nuclei, and the supernatants were centrifuged at 100,000 x g for 90 minutes in a Beckman model L5-65 ultracentrifuge. The supernatants were applied to a 8 ml column of DNase-S200 and a 20 ml column of fetuin-A15M in tandem at 4°. After washing the fetuin-A15M column extensively with buffer B, bound material was eluted with 150 mM /3-lactose in buffer B. Fractions were pooled, concentrated by ultrafiltration using a PM 10 membrane (Amicon) and frozen at -70°. Approximate yield was 50-100/~g per 109 cells. For iodination, 10/Jg of isolated protein was labeled with 1 mCi Bolton-Hunter reagent (NEN). 1Baum, LG and P Cresswell, unpublished results. 1080
Vol. 173, No. 3, 1 9 9 0
BIOCHEMICAL AND BIOPHYSICALRESEARCHCOMMUNICATIONS
Gel electrophoresis: One dimension SDS-PAGE was performed under reducing conditions (0.1% (v/v) 2-mercaptoethanol) as described by Maizel (15). Two dimension SDS-PAGE, with the addition of St. aureus V-8 protease (20/Jg/ml, Sigma) in the second dimension, was performed essentially as described by Cleveland (16). All SDS-PAGE gels were 10.5% (w/v) acrylamide. Isoelectric focussing was performed over a gradient of 3-10 according to the method of O'Farrell (17) modified so that focusing was performed on a 10 cm slab gel using a 2117 Multiphor slab gel apparatus (LKB). Standard proteins (Pharmacia) were run parallel to the sample. For charge-shift electrophoresis (18), 10/JI samples were applied to wells in the middle of 7.5 cm 1% agarose slab gels, in the appropriate detergent-containing buffers. Samples were subjected to electrophoresis for 80 minutes at 30 volts. Results Detergent extracts of one T and two B lymphoblastoid cells were applied to an affinity matrix of fetuin-A15M. Elution with 150 mM/3-lactose yielded the material shown in Figure 1, while incubation with the following saccharides were ineffective in eluting any material bound to the fetuin-A15M: galactose, N-acetyl galactosamine, mannose, glucose, N-acetylglucosamine, and xylose. In all three cell lines, the major carbohydrate binding protein eluted by /3-1actose has a molecular weight of 32,000 (CBP 32) by SDS-PAGE analysis. In addition, proteins of Mr = 29,000 (Swei, CEM) and Mr = 25,000 (TRal) are variably seen; these are presumed to be proteolytic products of CBP 32 (see below). [35S]-Iabelled preparations often contain material of Mr = 45,000 (e.g., CEM), which appears to be actin, by pl and differential mobility under reducing versus non-reducing conditions, as compared with purified actin (data not shown). To determine whether 29 kD material seen in Figure 1 is a proteolytic product of CBP 32, affinity purified unlabelled material containing both species was first subjected to
44> 40>
14~
CEM
Swei
TRal
Figure 1. Affinity purified CBP 32 from CEM(T), Swei(B) and TRaI(B) lymphoblastoid cells. Detergent extracts of [35S] labelled (CEM) or unlabeled (Swei,TRal) lymphoblastoid cells were applied to a fetuin-A15M column and eluted with 150 mM /3-lactose. Swei and TRal lanes were silver stained. 1081
Vol. 173, No. 3, 1990
BIOCHEMICAL AND BIOPHYSICALRESEARCH COMMUNICATIONS
67~,
B
44~. 40~..
14~
®
®
,23
Figure 2. The 32 kD and 29 kD species isolated by affinity chromatography appear to be related. Unlabelled material isolated from Swei cells was applied to 10 cm tube gels and subjected to SDS-PAGE. The tube gels were applied to slab gels after equilibration in sample buffer with or without the addition of 2/~g/ml S. aureus V-8 protease. Slab gels were silver stained. The appearance of only one major spot after digestion suggests that the protein has a protease-resistant core. Panel A, no protease added. Panel B, 2 pg/ml protease added to the sample buffer. Figure ~1"251]labeled CBP 32 isolated from Swei cells was reapplied to fetuin A15M. Both CBP 32 and the 29 kD proteolytic product were re-eluted with 150 mM /~-Iactose (lane 1) but neither 2 mM dithiothreitol in buffer B (lane 2) or 5 mM EDTA in buffer B (lane 3) eluted bound material from the affinity matrix.
electrophoresis on SDS-PAGE tube gels, and then applied to SDS-PAGE slab gels in the presence of St. aureus V-8 protease (20 pg/ml) (Figure 2). The peptide fragments from the 32 kD(a) and 29 kD(b) bands resolved on silver staining appear to be related. Both polypeptides were digested to one major spot (a' derived from a), approximately 8 kD less than the undigested precursors, and one minor spot (b, derived from b). The fragments derived from the 32 kD and 29 kD bands have the same spatial relationship on the gel as the intact molecules. This implies that both molecules lose the same peptide upon proteolysis. The 25 kD protein seen in Figure 1 appears primarily after 1082
Vol. 173, No. 3, 1990
BIOCHEMICALAND BIOPHYSICAL RESEARCH COMMUNICATIONS
i~i!i!i!i!~ li !,!i ii!i~~!
81x5_ 84,5 8.15 -
~
+
735 6.85
635-
Q
stds
CBP 32
®+
~
2
3
4
Figure 4. Isoelectric focussing of CBP 32. The sample was applied to the middle of a 3:10 gradient slab gel. Proteins were detected by Coomassie Blue staining. Isoelectric point standards: lentil lectin, basic band, 8.65; lentil lectin, middle band, 8.45; lentil lectin, acidic band, 8.15; horse myoglobin, basic band, 7.35; horse myoglobin, acidic band, 6.85; human carbonic anhydrase, 6.55. Figure ['1"251]labeled CBP 32 isolated from Swei cells was subjected to charge-shift electrophoresis. In sodium deoxycholate, CBP 32 migrated towards the anode (+); in cetyltrimethyl ammonium bromide, CBP 32 migrated towards the cathode (-); in Triton X100, CBP 32 did not migrate from the origin. Lane 1, CBP 32; lane 2, soluble F(ab, )2 fragments of rabbit immunoglobulin; lane 3, glycoproteins from TRal (mostly integral membrane proteins); lane 4, purified HLA class II molecules (a heterodimer of integral membrane glycoproteins). repeated freeze-thaw cycles; the irregular appearance of this product has prevented similar analysis. Unlike the soluble (S-type) lectins (5,19), the binding activity of CBP 32 did not appear to be thiol dependent, since the inclusion of reducing agents in extraction and elution buffers did not affect yields (data not shown). Figure 3 shows that the addition of dithiothreitol does not dissociate bound CBP 32, indicating that maintenance of an active binding site does not depend on intramolecular disulphide bonds. In addition, divalent cations do not appear to be essential for binding, as the presence of EDTA does not dissociate bound lectin (Figure 3). Native CBP 32 appears to exist as a monomer, since the protein could not be cross-linked by the bifunctional cross-linking reagent dithiobis (propionic acid N-hydroxysuccinimide ester) (data not shown). The isoelectric point of affinity-purified material was determined over a pH range of 3-10 (Figure 4). The proteins resolved into 3 distinct bands. This material was the 1083
Vol. 173, No. 3, 1 9 9 0
BIOCHEMICALAND BIOPHYSICALRESEARCHCOMMUNICATIONS
identical preparation used in Figure la (Swei); on the basis of staining intensity, the major band at 8.2 corresponds to the 32 kD protein. The two minor bands, at 8.5 and 8.7, thus represent the 29 kD and 25 kD material seen in Figure 1. This relationship suggests that the 25 kD protein is, like the 29 kD protein, a proteolytic product derived from CBP 32. While some soluble proteins may bind detergent, the majority of proteins capable of interacting with detergents are integral membrane proteins, which associate with detergent molecules via their hydrophobic transmembrane regions. To investigate whether CBP 32 is an integral membrane protein, the ability of the protein to bind detergent was examined by charge-shift electrophoresis. As shown in Figure 5, the direction of CBP 32 migration in 1% agarose gels changes according to the charge of the detergent used in the buffer solutions. In the top panel, CBP 32, equilibrated in the anionic detergent sodium deoxycholate, migrates towards the anode, while, in the middle panel, CBP 32 migrates towards the cathode in the presence of cetyltrimethylammonium bromide, a cationic detergent. CBP 32 does not migrate from the loading point in the neutral detergent Triton X-100 (bottom panel). These results indicate that CBP 32 has the ability to bind detergent, and thus is likely an integral membrane protein. Discussion Based on the previous observation that binding of glycoprotein micelles to activated lymphocytes and to B and T lymphoblastoid cells was inhibitable by -galactosides, we examined detergent extracts of these cells for carbohydrate binding proteins. Affinity chromatography of cell extracts on fetuin-A15M yielded one major carbohydrate binding (CBP 32), as well as two smaller proteins which appear to be proteolytic products of CBP 32. While a number of galactose specific lectins have been described in a variety of mammalian tissues, many of these are S-type lectins (5,19,20); CBP 32 differs from this class of lectins in that binding is not thiol dependent. In addition, CBP 32 appears to be an integral membrane protein, based on charge-shift electrophoresis studies. Indeed, attempts to extract CBP 32 with lactose in the absence of detergent were unsuccessful. CBP 32 may have similarities to the other major class of carbohydrate binding proteins, the C-type lectins, which are a heterogeneous group of soluble and membrane bound proteins with a wide range of saccharide specificities, yet which share a common carbohydrate recognition domain (5,7). However, unlike the C-type lectins, CBP 32 activity does not appear to be calcium dependent, as the presence of calcium was not necessary for purification, nor did EDTA elute bound lectin from fetuin-A15M. Further characterization of CBP 32 will allow more extensive comparison with other C-type lectins. 1084
Vol. 173, No. 3, 1 9 9 0
BIOCHEMICAL AND BIOPHYSICALRESEARCHCOMMUNICATIONS
The role of CBP 32 in human lymphoblastoid cells is not known. Lectin-sugar interactions appear to be important for a variety of human and murine lymphocyte functions, such as homing to lymphoid tissue (7,8), localization and maturation of thymocytes (4,6,9) and T cell activation (21,22). The initial finding that glycoprotein micelles bind only to activated and transformed lymphocytes, but not to resting cells, suggested that a galactose specific lectin may be involved in immune responses (13). Additional studies will reveal whether the novel lectin described here, CBP 32, is involved in interactions among activated and transformed cells. Acknowledgments The authors thank Alan Payne for photography, and Janet Railsback for manuscript preparation. Re, fences 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
Hooghe, R.J., and Pink, J.R.L. (1985) Immunology Today 6, 180-181. Coombe, D.R., and Rider, C.C. (1989) Immunology Today 10, 289-291. Stoolman, L.M. (1989) Cell 56, 907-910. Levi, G., and Teichberg, V.I. (1985) Biochem J. 226, 379-384. Drickamer, K. (1988) J. Biol. Chem. 263, 9957-9560. Sharon, N., and Lis, H. (1989) Science 246, 227-234. Laskey, L.A., Singer, M.S., Yednock, T.A., Dowbenko, D., Fennie, C., Rodriguez, H., Nguyen, T., Stachel, S., and Rosen, S.D. (1989) Cell 56, 1045-1055. Geoffroy, J.S., and Rosen, S.D. (1989) J. Cell Biol. 109, 2463-2469. Levi, G., and Teichberg, V.I. (1983) Immunol. Lett. 7, 35-39. Lee, R.T., Ichikawa, Y., Allen, H.J., and Lee, Y.C. (1990) J. Biol. Chem. 265, 7864-7871. Keida, C.M.T., Bowles, D.J., Rand, A., and Sharon, N. (1978) FEBS Lett. 94, 391-396. Decker, J.M. (1980) Molecular Immunology 17, 803-808. Apgar, J.R., and Cresswell, P. (1982) Eur. J. Immunol. 12, 570-576. Axen, R., Porath, J., and Ernback, S. (1967) Nature 214, 1302-1304. Maizel, J.V. (1971) Methods in Virol. 5, 179-245. Cleveland, D.W., Fischer, S.G., Kirschner, M.W., and Laemmli, U.K. (1976) J. Biol. Chem. 252, 1102-1106. O'Farrell, P.H. (1975) J. Biol. Chem. 250, 4007-4021. Helenius, A., and Simons, K. (1977) Proc. Natl. Acad. Sci. 74, 529-532. Barondes, S. (1984) Science 223, 1259-1264. Sparrow, C.P., Leffler, H., and Barondes, S. (1987) J. Biol. Chem. 262, 73837390. Hart, G. (1982) J. Biol. Chem. 257, 151-158. Piller,F., Piller, V., Fox, R.I., and Fukuda, M. (1988) J. Biol. Chem. 263, 1514615150.
1085
